Fig. 1

Overview of data processing steps. A total of seven datasets were retrieved from the public data registries (GEO and ArrayExpress). Four datasets were selected for integrated analysis, including samples from 30 patients with Sjögren’s syndrome (SjS) and 23 normal controls, covering 17,479 genes. The merged dataset was normalized using quantile normalization, and its batch effect was further corrected. Filtration of differentially expressed genes (DEGs), gene-set enrichment analysis (GSEA), unsupervised clustering, and key driver analysis (KDA) were performed according to the established methodology, and the clinical and molecular implications of the results were explored